RPA and AI: Dynamic solutions for insurtechs
The world of robotics is a dynamic and fast-moving place.
Just the very term conjures up sci-fi visions of a future managed by metallic, systematic creations that may or may not be advantageous to humans.
Controversy on the subject is rife. Elon Musk has often criticised the direction of robotics and recently voiced concern about the advancement of AI, stating that the current trajectory of the industry is worrying. In February 2020, he said, “Robots will be able to do everything better than us… I am not sure exactly what to do about this. This is really the scariest problem to me.”
However, that same concern has done little to curb Musk’s advances in self-driving cars and other dynamic projects, and therefore one must conclude that the advantages of robotics outweigh the consequences for the meantime at least.
RPA in insurtech
But whether you love or hate the idea of man versus machine, robotics technology is here to stay, and the arguments for its continued use and developments are persuasive. For example, within insurtech, robotic process automation - or RPA is an established technology that has been disrupting the insurtech sector for quite some time.
“It’s a mistake to think of RPA as a new technology,” says Jean-Marc Boxus, Associate Director at BCG Platinion, a South Africa-based IT solutions provider of AI, ML and RPA technology. “In reality the seeds are well-established, emerging from simple testing automation tools and, over time, made more mature by large scale outsourcers seeking to reduce costs and improve service levels.”
Simply put, the term RPA refers to software that instructs a computer to execute repetitive, everyday tasks which were previously carried out by humans. The idea is that if the machine handles the donkey work, the humans can get on with what they do best, namely innovating, creating and brainstorming.
The technology has proven popular and data shows the uptake has been high. “Studies from over three years ago anticipated that more than 70% of companies have tried some form of RPA technology to automate processes,” says Boxus. “However, these same studies also highlighted that only a very small portion of these companies (less than 5%) managed to scale the use of RPA to any kind of significant size – getting above, say 50 robots for example, presented a very real challenge.”
The reason? RPA just isn’t agile enough. Human cognicity plays a role in most processes. This is because of our skills to read and make sense of uncorrelated data. “An RPA solution does not offer capabilities to read text and unformatted data. Additionally, the technology is based on linear programming – RPA is based on simple rules, and therefore, cannot replace human judgment,” he stresses.
However, though the current scope is not as broad as it could be, RPA is still extremely effective. Richard Stewart, CEO and co-founder of the insurtech Untangl says the technology is a financially cost-effective and practical solution for companies because its main purpose is the automation of repetitive tasks, driven by the need to reduce operational costs or increase efficiency.
“Typically, these tasks will be relatively simple administrative processes where the rules are known and no decision making or machine learning is required,” he says.
RPA vs humans
Currently, RPA takes a repetitive task away from a human worker, but most experts agree that it is not capable of replacing a human staff member as it lacks the ability to discern. Cathal McGloin, CEO of the conversational AI platform company, ServisBOT which has helped Chill Insurance and The AA Ireland to rapidly create chatbots for customer engagement, believes RPA is still an essential tool for businesses today.
However, McGloin says that as valuable as automated systems are, they are currently no replacement for a human when it comes to customer service queries. “Customer onboarding and customer service bots lower the operational costs associated with a growing business. Some customer journeys can lend themselves to complete automation by bots, but they must always allow handover to a human if a customer query is complex or unforeseen. You don’t want your customers to get trapped in a bot loop.”
Stewart agrees. He also says RPA automation is as much about efficiency as it is about security - a major issue in these digitised times. "Looking at the potential IPO valuation of UI Path, a popular RPA vendor considered to be a major disruptor, you'd have to say it's on an upward trajectory and that adoption is bound to grow. RPA won't solve every business process challenge but it can take on the burden of admin, crucial for freeing up people to do more meaningful work."
He points out, "Many data issues are caused by human error or not implementing data protocols correctly, so using RPA could deliver a consistent approach to data handling and recording which is likely to be more easily vetted and audited."
AI, ML and RPA
With so many intelligent solutions to manage work processes, it is critical to define which technology does what, and which will be helpful in streamlining operations.
RPA is different to AI and ML solutions because it doesn't learn, it simply executes, says Stewart. "RPA manages simple, repeatable processes that can be defined. Think of it like working to a strict recipe. ML introduces a new cognitive layer where processes can be improved without human input: decisions to change are made by contextual reasoning, typically based on large amounts of data."
Security is a critical component in choosing software that creates a streamlined approach. Karli Kalpala, head of business and services design at Digital Workforce, an intelligent automation services provider based in Helsinki, agrees that the technology makes processes safer because it removes the human factor.
He says, “RPA has increased compliance, especially in the financial sector, it allows a complete audit trail of all the robot's actions and activities. Also, an automated process always performs the process through predetermined rules, making the process more predictable and less prone to human errors.”
Hyperautomation
But not everyone believes in the RPA magic bullet business model. In fact, some technology experts say its heyday may have passed and that in the UK particularly, innovation is needed to reinvigorate the market.
Faisal Abbasi, Managing Director Western Europe and MEMA of Amelia, the world’s largest privately owned AI firm, explains, “The RPA market is currently stagnating in the UK as trends in the automation industry have moved towards hyperautomation – intelligent, sentient business management software that can automate processes in a way that is significantly more impactful than standalone automation technologies.
He continues, “The most successful enterprise implementations of hyperautomation are those where solutions are built to mirror human intelligence: combining digital emotional intelligence with natural language understanding.”
Abassi says that new technology is in increasing demand from a broad range of sectors. “Currently, the largest sectors that stand to benefit from automation in the UK include insurance, banking and retail organisations. With huge customer-facing responsibilities, these businesses are frequently handling millions of calls and customer requests a month.
“Right now, we’re seeing an increase in companies such as those looking to advance automation in order to help them reduce the manpower required to handle basic, everyday enquiries, including requesting claims and customer information or sending a bank statement.”
Combined technologies
Furthermore, the differences between the three technologies will eventually disappear as combined solutions look set to trend within the next five years, points out Kapala. “The gap between RPA and AI-systems is rapidly closing. The major providers are introducing the possibilities of AI subsystems to the RPA world. Many standard tools like OCR or Google Search are already using advanced AI which might not be evident to the user.”
McGloin concurs. He says, “Automated conversations are transforming business processes, driving increased speed and convenience and reduced friction and costs. As this trend develops, we see the line between conversational AI-based chatbot technology and robotic process automation (RPA) increasingly blurring.
“A chatbot understands and simulates human conversation, while an RPA robot emulates human actions. Natural language processing (NLP) plays a role within both technologies, chatbots interpret conversations from voice or text channels, while RPA bots extract language and data from documents, files, forms and browsers.”
With cutting edge technology amalgamating, RPA merging with artificial intelligence (AI) and machine learning (ML) will result in superbots that offer hyper-efficient learning solutions that could, theoretically replace human workers.
Kalpala says, “RPA is now evolving towards hyperautomation where RPA bots are augmented with AI and ML components to allow more human-like features, thus allowing the more complex end to end processes to be automated. Gartner has quoted hypeautomation to be the most important technology trend of 2020.”
He continues, “World economic forum chairman and founder Claus Schwab has said that intelligent automation a.k.a hyperautomation is the key enabler of industry 4.0 the fourth industrial revolution, ‘automation of knowledge work’.”
Kalpala adds, “This introduction and integration are, however, still ongoing. Furthermore, there are specific apparent differences on how classical RPA methods and AI-based methods are used, so it will remain essential to understand, at least on some abstract level, the differences between these two types of technology.”
Trending technologies for insurtech
- Artificial intelligence (AI) refers to machines that are programmed to think like humans and mimic their actions. The term can also be used with regard to any machine that exhibits human mind traits, including learning and problem-solving.
- Machine learning (ML) is a form of AI that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it to learn for themselves.
- Robotic Process Automation (RPA) is a general-purpose integration and process automation technology. Its main task is to automate the kind of processes typically undertaken by workers. Jobs that are repetitive can be taken over by computer systems, guaranteeing that they will be fulfilled without human error. RPA, unlike AI and ML, only carries out tasks it is programmed to manage, and does not learn.
- Hyperautomation is the amalgamation of three primary technologies, namely AI, ML and RPA as well as Process Mining (MP) to create digital workers and automate processes that are significantly more productive than traditional automation solutions.
RPA security
With cyber breaches happening to the world’s biggest technology giants, the threat of attack is very real for all businesses globally. Insurtechs are more vulnerable than most, because the data they hold for customers is often highly sensitive.
RPA is one part of the process that keeps such information secure, says Jean-Marc Boxus, Associate Director at BCG Platinion. “With RPA data is indeed safer,” he says. “Ask a robot to input data and it will do so, systematically and until the programme is compete; ask a human to log all their actions in a registry and you can bet that after just a few hours errors and omissions will creep in. Robots do not make processing mistakes. they do not slip up doing simple copy-and-paste tasks – something we humans cannot claim.”